PURPOSE: Cardiac rehabilitation (CR) provides effective secondary prevention services, but many patients fail to complete the recommended program. The purposes of this study were to describe completion rates in a hospital-based outpatient CR program, and to identify factors associated with patients failing to complete CR because of nonmedical and medical reasons. METHODS: Data used for the analyses were from a hospital-based CR program involving 526 discharged patients between January 1996 and February 2002. Patient discharge status was classified into three categories: complete, noncomplete-medical reasons, and noncomplete-nonmedical reasons. Logistic regression modeling identified factors associated with the groups failing to complete CR. RESULTS: The rate of CR completion was 58% (304/526). Among the 222 patients who did not complete CR, 139 (63%) had nonmedical reasons. As compared with the patients who completed CR, the adjusted odds ratio (AOR) for those who did not complete CR because of nonmedical reasons were more likely to be employed (AOR 2.2), to be obese (AOR 2.5), to be smokers (AOR 2.1), and to have shorter 6-minute walk distances (AOR 1.7). They were less likely to be women (AOR 0.6) or have diabetes (AOR 0.5). Patients not completing CR for medical reasons were more likely to be categorized as being at high clinical risk (AOR 4.2) and having shorter 6-minute walk distances (AOR 1.9). CONCLUSION: Except for low functional capacity, baseline factors associated with patients failing to complete CR differed on the basis of medical or nonmedical reasons. The development of interventions that address the special needs of patients with low functional capacity may be especially important in attempts to retain this high-risk group in CR therapy.